Neural tube defects ( Ch. 40.8 )

The diagnosis of anencephaly is straightforward: the cranial vault cannot be visualised in the standard view for biparietal diameter measurement. Detection of open myelomeningocele is more complex. Although larger defects may be suggested by gross disruption in vertebral integrity in the longitudinal plane or by soft-tissue signs, smaller defects will only be apparent in the horizontal planes of a few localised vertebrae, as subtle splaying in the lateral processes. These views can be difficult to obtain, especially if the fetal spine lies against the uterine wall, and in this context screening has been greatly facilitated by two cranial signs found in almost all fetuses with myelomeningocele. Scalloping of the frontal bones gives the head a lemon-shaped appearance (‘lemon’ sign), whereas the normally dumbbell-shaped cerebellum appears either absent or banana-shaped (‘banana’ sign: Fig. 9.2 ) ( ; ; ). The latter results from downward herniation of posterior fossa contents, and the former from the subsequent reduction in intracranial volume.

(A) The normal dumbbell shape of the cerebellum (arrows) on ultrasound at 18 weeks. (B) Anterior curvature of the cerebellum, the ‘banana’ sign (arrows), in a fetus of similar gestation with open spina bifida.

Ultrasound screening using the lemon and banana signs should theoretically detect 96–100% of myelomeningoceles ( ; ). Ultrasound features such as head circumference greater than the 90th percentile and lesions at L3 and above have been been demonstrated in multivariate analysis to be independently associated with lower rates of survival in children born with spina bifida ( ). In a long-term follow-up study of an unselected cohort of infants born with spina bifida between 1963 and 1971, by the mean age of 30 years the survival rate was approximately 50%. Only one-third of the survivors lived independently, with up to a third requiring daily care. Sensory levels recorded at infancy were correlated with rates of mortality and morbidity, with higher sensory levels associated with poorer outcome ( ; ). See Chapter 40 part 8 for more information on neonatal aspects of spina bifida.

It is important to recognise that ultrasound is both a screening and a diagnostic test, and that small spinal lesions in fetuses with suspicious cranial signs on screening may only be detected on detailed scanning by a very experienced operator, and that the antenatal prediction of the level of the lesion may not always correspond to that found after birth. In cases of genuine uncertainty, an elevated amniotic fluid concentration of acetylcholinesterase may also be measured to confirm the presence of an open NTD ( ).

Other craniospinal malformations

Ventriculomegaly is diagnosed in utero by elevated ratios of various measurements of the lateral ventricle to hemispheric width in the transverse plane ( ; ) or enlargement of the atrium of the posterior horn of the lateral ventricle ( ). Unless there is progressive or gross dilatation, caution should be exercised in the interpretation of mild ventriculomegaly, especially in the mid-trimester. In a recent review of the literature on isolated mild ventriculomegaly, defined as ventricular atrial width of 10.0–15.0 mm, there was no conclusive evidence that the width of the ventricular atria was associated with the risk of neurodevelopmental outcome ( ). The pooled incidence of neurodevelopmental delay in this review was 11%. Limitations in the current literature include evidence from small retrospective case series and lack of long-term follow-up studies. Important adverse prognostic indicators include progressive ventricular dilatation and other associated abnormalities.

The level of obstruction is determined by examining the third and fourth ventricles and the aqueduct of Sylvius. Hydranencephaly is distinguished from severe hydrocephalus by the absence of midline structures and the lack of a residual cortical rim, but the distinction may be difficult in extreme cases. In holoprosencephaly, the extent of midline ventricular fusion varies with the degree of failure of cleavage of the prosencephalon ( ), and there are often concomitant facial anomalies. The diagnosis of microcephaly should only be made in the presence of serial measurements of head circumference 3–4 standard deviations ( sd ) below the mean, so as to exclude growth retardation or incorrect dating. The ultrasonic appearances of encephalocele, and intracranial cysts, tumours and haemorrhage, are well described. Agenesis of the corpus callosum, increasingly recognised as separation of the lateral ventricles with upward displacement of the third ventricle, is more difficult to diagnose, and is often detected in association with other CNS abnormalities ( ).

The diagnosis of posterior fossa abnormalities, and most notably the Dandy–Walker malformation and its variants, is challenging. The classic appearance is of complete or partial agenesis of the cerebellar vermis with a posterior fossa cyst ( Fig. 40.90 ). However, in one series from a tertiary unit in the UK, the correlation with postmortem was only 43% (6/14) for this abnormality ( ). This has led to evaluation of fetal MRI for complex neurological abnormalities (see below).

Cardiac defects

Following characterisation of the normal ultrasonic appearances of the fetal heart, a wide range of defects have been diagnosed. Inspection of the four-chamber view in a transverse plane during the routine 18-week scan detects approximately 20–40% of severe congenital heart disease ( ; ). This view is abnormal with major defects such as hypoplastic ventricles, atrioventricular canal defects and tricuspid atresia, although minor lesions such as septal defects may be missed. Visualisation of venoatrial and ventriculoarterial connections is more complex, but if it were introduced into routine screening it would theoretically increase the detection rate to greater than 80% ( ). Some cases are missed because of the evolution of the abnormality such that the heart may appear to be normal at the time of screening. The introduction of the five-chamber view, sagittal and parasagittal views has increased the detection rate of outflow tract anomalies ( ; ; ).

Indications for fetal echocardiography include an abnormal four-chamber view or outflow tracts, an affected sibling or parental history, diabetes, exposure to cardiac teratogens, raised nuchal translucency (NT) measurement in a euploid fetus, and monochorionic twin gestations. A high degree of accuracy can then be achieved for diagnosis of major vessel lesions such as pulmonary stenosis, truncus arteriosus and transposition of the great vessels ( ). Colour flow Doppler facilitates the demonstration of cardiac structure, and M-mode and pulsed-wave Doppler provide an index of cardiac function, which is particularly useful in arrhythmias.

Intrathoracic defects (for neonatal surgical management, see Ch. 29.4 )

Most intrathoracic lesions identified prenatally are benign, but their significance is the association with pulmonary hypoplasia. Indirect indices of the severity of compression on the developing lung are the degree of mediastinal shift, the presence of polyhydramnios due to limited swallowing, and hydrops secondary to obstructed venous return. Left-sided diaphragmatic hernias are detected on ultrasound because of displacement of the heart to the right, fluid-filled bowel within the chest, and absence of the stomach intra-abdominally ( ; Nakayama et al. 1985). As herniated liver has the same echotexture as the lung, right-sided lesions are more difficult to detect, especially in the absence of polyhydramnios or pleural effusion. Clues include derangement in the normal intra-abdominal course of the gallbladder and intrahepatic vein, although small lesions may go undetected.

The appearances of congenital cystic adenomatoid malformation (CCAM) (now often termed congenital pulmonary airway malformation; Ch. 27.5 ) of the lung vary from solitary large cysts to solid echogenic lesions, with the worst prognosis ( Fig. 9.3 ) ( ; ; Adziek et al. 1985). The assessment of prognosis is difficult, although in more recent series it appears good, other than in a small minority (6%), in which both mediastinal shift and hydrops are present ( ; ). Some disappear, many get smaller, and only a few stay the same or progress. Diagnosis cannot always be certain antenatally, and differentiation between CCAM, sequestrated lung, tracheal or bronchial atresia, and congenital diaphragmatic hernia may be difficult ( ; ). Mediastinal teratomas have features on ultrasound similar to solid type III or microcystic CCAM lesions ( ).

Congenital cystic adenomatous malformation (CCAM) of the lung. A longitudinal view of a large macrocystic CCAM.

Gastrointestinal defects (for neonatal surgical management, see Ch. 29.4 )

Oesophageal atresia can be diagnosed when polyhydramnios occurs with non-visualisation of the fetal stomach, although these signs may not be present in the common form associated with tracheo-oesophageal fistula ( ). Duodenal atresia produces polyhydramnios and a characteristic ‘double-bubble’ appearance ( Fig. 9.4 ), which may not be apparent until the third trimester ( ). Associated anomalies are common in the above two conditions, unlike in more distal obstructions. Small-bowel obstructions are more likely to be associated with increased amniotic fluid volume than are large-bowel obstructions such as anal atresia, which may go undetected in utero. Peristalsis may be seen ( ) and bowel perforations may show up as ascites or, more commonly, hyperechogenicity from meconium peritonitis ( ). Caution must be exercised when gut echogenicity is detected as an isolated finding, as, although this may be associated with cystic fibrosis ( ) and aneuploidy, the majority occur in normal fetuses ( ). Meconium peritonitis can also produce pseudocysts, the differential diagnosis of which includes gastrointestinal duplications and choledochal, mesenteric and ovarian cysts.

The ‘double-bubble’ appearance of duodenal atresia.

The distinction of exomphalos from gastroschisis is crucial, given the high incidence of cardiac and chromosomal abnormalities in the former ( ). Aneuploidy seems largely confined to fetuses in which the herniated liver is not present within the exomphalos ( ; ). More severe degrees of failure of fusion of the ectomesodermic folds, such as ectopia cordis, ectopia vesicae, pentalogy of Cantrell and the body stalk anomaly, are readily apparent. In contrast, gastroschisis is rarely associated with aneuploidy or genetic conditions. Non-genetic associations include young maternal age, socioeconomic deprivation, ethnicity, maternal smoking, medication (aspirin, ibuprofen, paracetamol and decongestants), illicit drug use and maternal nutritional status ( ). There are however trends in the incidence of gastroschisis which remain unexplained. These include the increasing incidence of gastroschisis in the UK and the clustered distribution of cases in the population ( ; ). These patterns suggest a clustered exposure to one or more risk factors which are increasing over time ( ).

Genitourinary defects (for neonatal diagnosis and management, see Ch. 35.2 )

Renal and urinary tract abnormalities are common and comparatively easy to detect, largely because obstructive lesions manifest as cystic spaces, whereas those with poor urine output are characterised by oligohydramnios. Major anomalies such as renal agenesis or low obstructive uropathy will be detected on routine scan at 18–20 weeks, when urine output makes a major contribution to amniotic fluid volume, whereas more minor lesions, such as mild ureteropelvic junction obstruction, may not be obvious until later. As the lack of amniotic fluid in renal agenesis significantly impairs the ultrasound picture, it can be extremely difficult to demonstrate the absence of kidneys in the renal fossae. In these circumstances, referral for confirmation by transvaginal ultrasound and/or amnioinfusion has been recommended ( ). However, colour flow Doppler imaging of the renal arteries may be adequate to provide the diagnosis: the absence of renal artery colour image confirms no functioning tissue ( ).

Multicystic kidneys are distinguished from hydronephrotic kidneys by their cystic spaces being more peripheral and variable in size, and their stroma more central ( Fig. 9.5 ) ( ). The cysts of infantile polycystic kidneys are too small to be resolved by ultrasound, but the kidneys appear enlarged with abnormal echogenicity, associated with oligohydramnios. Occasionally with later-onset infantile polycystic kidney disease the kidneys may appear normal in utero ( ).

Multicystic kidney disease.

The significance of mild pelvicalyceal dilatation remains controversial. Studies show that progressive enlargement, or an anteroposterior diameter of >10 mm in the third trimester, is more likely to be associated with pathology ( ). The ultrasound picture in low obstructive uropathy depends on the severity and duration of obstruction. The bladder is variably enlarged and thick-walled, and careful scanning reveals dilatation of the upper urethra in those with posterior urethral valves (PUVs) ( ). Oligohydramnios, upper tract dilatation and hyperechogenic fetal kidneys may also be present.

Skeletal defects (for neonatal diagnosis and management, see Ch. 31 , Ch. 34.4 )

Isolated malformations detected on routine scanning include kyphoscoliosis, hemivertebrae, limb reduction and shortening, sacral agenesis, polydactyly and flexion deformities. Over 100 distinct skeletal dysplasias are amenable to prenatal diagnosis, both by serial measurement of long bones and by detection of abnormal skeletal shape or mineralisation. Although severe limb shortening, abnormal head or chest shape, or polyhydramnios may alert the sonologist to their presence, determination of the exact type of skeletal dysplasia is difficult in the absence of a previous history, and requires detailed evaluation of hands and feet, thoracic dimensions, face and cranium, and measurement of all the long bones, before consulting comprehensive tables of diagnostic features ( ). Even then, diagnosis may only be made postnatally, when additional investigations such as skeletal X-ray are available. In achondrogenesis, thanatophoric and diastrophic dwarfism, severe limb reduction will be obvious by 18 weeks ( ), whereas in the heterozygous form of achondroplasia this may not be observed until almost the third trimester ( ). If achondroplasia is suspected, amniocentesis can be performed for DNA testing on fibroblasts to detect or exclude the fibroblast growth receptor 3 (FGR-3) mutation known to cause achondroplasia ( ). Abnormal bone shape is a feature of camptomelic and thanatophoric dysplasia, whereas fractures, callus formation and hypomineralisation may be seen in osteogenesis imperfecta types II–IV ( ). Hypomineralisation is also seen in hypophosphatasia and achondrogenesis. Radial aplasia may be associated with trisomy 18, but also with rare genetic syndromes such as Fanconi’s anaemia and the thrombocytopenia–absent radius syndrome.

Soft-tissue abnormalities

Cleft lip, whether isolated or associated with cleft palate, can be detected by imaging the fetal face in coronal and transverse views ( ). Rarer midline clefts are often accompanied by other midline defects, such as holoprosencephaly, ethmocephaly or proboscis. The diagnosis of isolated cleft palate is extremely difficult prenatally ( ). These lesions are usually only detected on detailed scans of patients who have either other abnormalities or an at-risk history. Isolated unilateral clefts are rarely associated with either aneuploidy or genetic syndromes; however, these risks increase with bilateral clefts and the presence of associated anomalies.

Cystic hygromas are readily apparent as multiseptate thin-walled cystic lesions, found most commonly around the dorsolateral region of the neck ( ). As they result from obstruction of the jugulolymphatic channels ( ), hygromas frequently progress in utero to hydrops and fetal demise ( ), although spontaneous resolution is possible ( ). They have a strong association with aneuploidy (50–80%), and therefore karyotyping is advised. Hydrops is obvious as generalised skin oedema with ascites, and in many cases there will also be pericardial and pleural effusions, placentomegaly and polyhydramnios.

3D/4D ultrasound

The use of three-dimensional (3D) ultrasound in prenatal diagnosis has increased in recent years with the development of computational technologies ( Fig. 9.6 ). In cases of facial anomalies, the use of 3D ultrasound has improved the accuracy of detection of facial clefts ( ). In CNS abnormalities, assessment of the corpus callosum is now possible using the C-plane of 3D volumes (VCI) ( Fig. 9.7 ), and simultaneous visualisation with 3D ultrasound may improve the precise location of NTDs.

Three-dimensional ultrasound of the face of a fetus with known cardiac abnormality and oesophageal atresia. No dysmorphic facial features were confirmed.

(A) Three dimensional volume contrast imaging (3D VCI) demonstrating the C-plane to visualise the corpus callosum, cerebellum and vermis – only the C-plane image shown. (B and C) 3D VCI showing A-plane image and C-plane volume reconstruction of the sagittal midline view. In this case there was complete agenesis of the corpus callosum, which was confirmed by magnetic resonance imaging.

Transvaginal ultrasound and MRI

The greater resolution provided by high-frequency transvaginal transducers for structures within their 6–7-cm focal zone is ideal for visualisation of the first-trimester fetus.

MRI of the fetus was first described in the 1980s. Recently, faster capture times have helped to counter the problems created by fetal movements ( Fig. 9.8 ). Although MRI has been used to aid the prenatal diagnosis of many conditions, its strength lies in assessing intracranial anatomy, and in particular the posterior fossa. Conditions including agenesis of the corpus callosum are reliably detected by MRI; the recognition of neuronal migration disorders depends on the severity of the disorder and the gestation at which the MRI is made. MRI has a place in prenatal diagnosis, although experience in this form of fetal imaging is still being collected.

Fetal magnetic resonance imaging. Sagittal view of a normal mid-trimester fetus.

Biochemical

The average maternal serum alpha-fetoprotein (AFP) value in Down syndrome pregnancies is 0.7–0.8 multiples of the median ( ). As this association is largely independent of maternal age, the two were then combined to give each woman a specific age- and AFP-adjusted risk ( ). Subsequently, raised human chorionic gonadotrophin (hCG) (Bogart et al. 1989), particularly the free β-hCG subunit, and low unconjugated oestriol ( ) levels were found to be associated with trisomy 21. Again, in the absence of any relation to maternal age or each other, their levels and maternal age were next combined using an algorithm to predict fetal risk ( ). The reason for these biochemical changes is not yet understood, but is thought to relate to functional immaturity, producing a delay in the normal gestational rise or fall. Cut-offs of 1 : 250–1 : 300 have shown a sensitivity of up to 70%, with a false-positive rate of 5%. The term ‘false-positive’ in this context describes the percentage of the population that would be given a high-risk result if the test were taken up by the entire population. Subsequent refinements using free β-hCG ( ) and a prior dating scan ( ), and confining the assay to 15–18 weeks’ gestation, have increased the accuracy of the test. Additional biochemical measurement of unconjugated oestriol (triple test) and with inhibin (quadruple test) improves the detection rate ( ). The quadruple test has a sensitivity approaching 80% for trisomy 21 with a fixed false-positive rate of 5% ( ). Detection rates are greater in older mothers (i.e. over 35 years of age), but this is at the cost of higher false-positive rates ( ).

As biochemical screening takes place in the mid-trimester, it does not allow susceptible women the option of first-trimester CVS for karyotyping. The drive now is to bring screening for aneuploidy into the first and early second trimester, to allow earlier detection and easier termination procedures to be performed.

Ultrasound plus biochemistry in first and second trimesters

There have been major advances in ultrasound screening for aneuploidy. In the first trimester (10–14 weeks) the association between increased NT ( Fig. 9.9 ) and aneuploidy has been confirmed in several studies ( ; ; ). The association is with trisomies 21, 18 and 13 in particular. A sensitivity, when related to crown–rump length and maternal age, of almost 80%, with a false-positive rate of 5%, has been reported by one group in a low-risk population cohort of 96 127 ( ).

Transabdominal ultrasound view of a nuchal translucency measurement in a fetus at 12 weeks’ gestation.

Recently, NT measurement has been combined with two first-trimester biochemical markers to improve the efficacy of early screening for trisomy 21. Levels of free β-hCG are increased in the serum of mothers carrying a fetus with Down syndrome, while levels of pregnancy-associated plasma protein A are decreased. When this test is combined with NT, the combined test yields a detection rate of 90% for a similar false-positive rate ( ; ). The integrated test, which combines first-trimester screening (combined test) with second-trimester screening (AFP, oestriol and inhibin), further increases the detection rate of Down syndrome to 95% ( ).

Recent attention has focused on the relevance of the presence/absence of the fetal nasal bone ( ). In a series from King’s College Hospital, London, UK, the nasal bone was absent in 43 of 59 (73%) trisomy 21 fetuses and in 3 of 603 (0.5%) chromosomally normal fetuses. The likelihood ratio, therefore, for trisomy 21 was 146 (95% confidence interval (CI) 50–434) for absent nasal bone and 0.27 (0.18–0.40) for nasal bone being present. Subsequent work by the same group ( ) has shown in a retrospective series that the integration of nasal bone screening into the OSCAR test (One-Stop Clinic for Assessment of Risk) would lead to a detection rate of 97%; alternatively, for a false-positive rate of 0.5%, the detection rate is 90.5%. There is some evidence that the detection rate increases when used in addition to the combined test ( ), although this has not been consistently reproduced by other groups ( ; ). Additional ultrasound markers to increase the sensitivity and reduce the false-positive rates are also being assessed, including the ductus venosus waveform and triscuspid regurgitation ( ; ).

Subtle markers of aneuploidy

With advances in ultrasound resolution, many structures not previously visualised, such as the digits, feet and soft tissues of the neck, can now be demonstrated. Accordingly, the minor malformations and abnormal postures characteristic of aneuploid neonates may be seen in utero. Postaxial polydactyly is found more frequently in trisomy 13 than in 18, but ventricular septal defects are common in both ( ). The profile of the trisomy 18 fetus reveals micrognathia and a protuberant upper lip ( ). The hands remain clenched in trisomy 18, with characteristic overlapping of the fingers, and the typical rocker bottom and equinovarus deformity are found in the feet. The sonographic features of trisomy 21 are more elusive. These may include increased nuchal thickness (first or second trimester), mild renal pelvic dilatation, hyperechogenic bowel, brachycephaly and hypoplasia of the middle phalanx of the fifth finger (clinodactyly).

Amniocentesis

Amniocentesis is the commonest invasive procedure for prenatal diagnosis. Most are done at 14–16 weeks’ gestation, when the amniotic cavity contains 150–200 ml of fluid, allowing 15–20 ml to be withdrawn without complication (working rule is x ml withdrawn = x weeks in gestation). A 22G needle is inserted transabdominally and guided to a pool under ultrasound control ( Fig. 9.10 ). Simultaneous ultrasound monitoring reduces the number of dry and bloody taps ( ; ) and obviates the rare risk of severe fetal trauma. It has thus replaced the older technique of ‘semiblind’ insertion following ultrasound identification of a pool. Transplacental insertions have been linked with an increased miscarriage rate ( ; ) and should be avoided. Even with an extensive anterior placenta, a small window avoiding the placenta can usually be found.

(A) Amniocentesis: the 22G needle is introduced under ultrasound guidance (from top left). (B) Chorionic villus sampling: the 18G needle is introduced under ultrasound guidance (from top left).

Patients are quoted a risk of spontaneous miscarriage attributable to the procedure of 1%, based on the results of the only randomised controlled trial (RCT; ). Amniotic fluid contains cells desquamated from fetal skin, gastrointestinal, urogenital and respiratory tracts, and the amnion. In view of their small number, up to 2 weeks’ cell culture is required prior to cytogenetic analysis, although with new techniques this period is shortening. A major disadvantage of amniocentesis is that termination of affected fetuses is not performed until well into the mid-trimester. Although amniotic fluid can be satisfactorily obtained and cultured at 11–13 weeks ( ; ), the risks are increased with a greater miscarriage rate than with CVS ( ), and therefore is not recommended.

Approximately 0.5% of amniotic cell cultures fail to grow, and maternal cell contamination leads to diagnostic difficulty in 0.2% ( ; ).

Fluorescent in situ hybridisation (FISH; Ch. 8 ) was introduced to allow rapid detection of trisomy 13, 18 and 21 together with sex chromosome aneuploidies. FISH allows the detection of specific DNA sequences with chromosome-specific painting probes.

Quantitative fluorescent polymerase chain reaction (QF-PCR), a more recent technique, is a rapid method of detection of trisomy 13, 18 and 21 and sex chromosomes. Compared with FISH, this technique is entirely automated, quicker, feasible on fewer cells and more cost-effective with comparable detection rates ( ; ; ). Maternal cell contamination is also much more readily identified ( ). At present in the UK, the National Screening Committee advocates the use of QF-PCR for the detection of trisomy 13, 18 and 21 in women with high-risk screening results but with normal ultrasound. Full karyotope is only advocated in cases with abnormal ultrasound findings (including NT ≥3 mm) or family/previous history of chromosomal abnormalities. Most undetected karyotype abnormalities as a result of this clinical policy are not clinically significant ( ).

Chorionic villus sampling

Although obtaining chorionic tissue suitable for cytogenetic and biochemical analysis was first reported more than 30 years ago ( ), CVS was only introduced into clinical practice in the mid-1980s. CVS was originally performed transcervically, but is now most commonly performed transabdominally ( Fig. 9.10 ). Safety appears to be better with the transabdominal route (miscarriage rates 1% versus 4%) ( ; ). Initially, CVS was performed between 8 and 12 weeks. However, concern was raised following a report of limb reduction defects, some in association with the oromandibular syndrome ( ). In all cases, CVS was performed before 66 days’ gestation, and by single-needle transabdominal aspiration. Subsequent population and case–control studies have been unable to confirm any link ( ), but theoretically it seems plausible that a procedure which may cause embolism, thrombosis or vasoconstriction at the time of limb bud formation may lead to such malformations. Thus, it has been recommended that CVS be performed after 10 weeks’ gestation ( ; ). CVS rates remain high owing to first-trimester combined screening leading to aneuploidy testing, and for cases requiring genetic molecular testing.

The CVS-related fetal loss rate before 28 weeks above the background rate is 1–2% in centres with much experience ( ). One problem with CVS is a 1.0–1.5% incidence of confined placental or pseudomosaicism, where a discrepancy exists between chorionic and fetal karyotypes, necessitating further investigation by amniocentesis ( ). In most cases, a bizarre aneuploid or polyploid mosaic is identified in the chorion, whereas the fetal karyotype, assessed from skin fibroblasts, is normal.

Fetal blood sampling

FBS is performed by direct ultrasound-guided needling of various fetal vessels. This is done as an outpatient procedure under local anaesthesia from 17 weeks’ gestation. Sedation is rarely necessary. The most common approach, which involves inserting a 20G needle transabdominally into the umbilical vein about 1 cm from the placental cord insertion ( ), yields an adequate sample in 97% of cases ( ). Maternal contamination from inadvertent intervillous sampling is ruled out before removal of the needle by comparing the sample’s MCV distribution, determined rapidly on a particle size analyser, with that of the mother ( ). Even at term, fetal MCV, which declines rapidly with gestation, remains significantly higher than that of healthy mothers ( ). The vein is the preferred vessel to sample, being simpler and safer ( ); accidental sampling of the artery can be confirmed by ultrasonic observation of the direction of flow following injection of sterile saline ( ). When there is difficulty approaching the cord insertion because of obesity, oligo- or polyhydramnios, or fetal position, blood may be aspirated from the intrahepatic portion of the fetal umbilical vein ( ; ). Although FBS from the intrahepatic vein is more difficult than at the cord insertion, it obviates the need for laboratory confirmation of its source, and in multiple pregnancies the operator is certain as to which fetus is sampled ( ).

Determining the loss rate attributable to FBS is difficult, as fetal demise in the weeks after the procedure may instead be due to the underlying high-risk indication. This was clearly demonstrated in one series, which showed increased loss rates in procedures performed on sicker fetuses, i.e. prenatal diagnosis (2%), fetal structural abnormality (6%), assessment of severe FGR (14%), and hydrops (25%) ( ). A reported summation of all the published series calculated an overall loss rate in procedures performed in low-risk cases of 2.7% ( ), although others have reported a lower loss rate of 0.9% ( ), and this figure has been corroborated by the US International Registry ( ). Such results are only achieved after considerable training and experience. Most losses are due to cord haematoma, cord tamponade or haemorrhage, and, unlike losses after CVS or amniocentesis, are apparent at the time of the procedure. In late pregnancy, emergency caesarean section is performed to salvage these infants, although some may be damaged. Intra-amniotic bleeding is observed ultrasonically after 40% of samplings ( ; ), and a histological study of cords within 48 hours of FBS showed that a degree of extravasation occurs in all cases ( ). This bleeding is almost always transient, owing to the abundance of thromboplastins in amniotic fluid ( ).

Skin and muscle biopsy

Prenatal diagnosis of many severe genodermatoses necessitates histological and ultrastructural examination of fetal skin, obtained at 18–22 weeks, initially by fetoscopy ( ) but more recently by ultrasound-guided techniques ( ). The usual site chosen is the fetal buttock or leg. In a review of 269 pregnancies at risk of severe inherited skin disease the main indication for fetal skin biopsy was the risk of recurrence of epidermolysis bullosa. In such families, in over 90% of cases a prenatal diagnosis was established ( ). Epidermolysis bullosa letalis is characterised by separation of the epidermis from dermis at the lamina lucida on light microscopy, and a paucity of hemidesmosomes on electron microscopy ( ) ( Ch. 32 ). The prenatal diagnostic features of epidermolysis bullosa dystrophica, epidermolytic hyperkeratosis, harlequin ichthyosis and Sjögren–Larsson syndrome have similarly been described ( ; ; ). In oculocutaneous albinism, in which there is a lack of active melanin synthesis in hair bulb melanocytes, the biopsy must be taken from a hair-bearing area such as the scalp ( ). In the review above, over a 25-year period the miscarriage rate for fetal skin biopsy was approximately 1% ( ). Fetal skin scarring was observed postnatally in only 6 of the 191 cases.

Although DNA analysis is available for the diagnosis of the most common muscular dystrophies, most notably Duchenne muscular dystrophy, there are circumstances when a direct fetal muscle biopsy is required. This is usually when genetic testing has proved inconclusive. The procedure is performed under ultrasound control with the biopsy forceps guided into the outer aspect of the fetal buttock so as to avoid major structures.

Anti-D prophylaxis

RhD-negative women are at risk of sensitisation from fetomaternal bleeding, not only at delivery but also in other situations, such as external cephalic version and amniocentesis. Although the prevention programme has reduced neonatal deaths attributable to haemolytic disease of the newborn (HDN) from 18.4/100 000 in 1977 to 1.3/100 000 in 1992, there remains a sensitisation rate of around 1.5% among Rh-negative women. Administration of 500 IU of anti-D at 28 and 34 weeks can reduce the risk of immunisation to 0.2% ( ). Anti-D prophylaxis is now recommended by the and the , with either a dose of 500 IU at 28 and 34 weeks, or a single larger dose of 1500 IU early in the third trimester. The rationale is that all Rh-negative women are at risk from hidden bleeds. The introduction of routine antenatal prophylaxis at 28 weeks in the UK has reduced the incidence of new sensitisation ( ; ).

It is, however, important to note that clinically significant disease can be caused by other red cell alloantibodies such as Rh-c, Kell, Rh-E and Fy ^a (Duffy). At present, immunoprophylaxis is not available for non-RhD disease and so immunisation will continue to occur.

Antenatal screening and management of RBC alloimmunisation

Routine serological testing of women is carried out:

• •
  

  to identify pregnancies at risk of fetal and neonatal alloimmune disease (HDN)

  
  
• •
  

  to identify RhD-negative women who require antenatal anti-D prophylaxis

  
  
• •
  

  to provide compatible blood swiftly in emergencies.

All women who have no antibodies at 10–16 weeks’ gestation should be tested once again between 28 and 36 weeks. Some workers believe that RhD-negative women should have two further tests, one at 28 weeks and one at 34–36 weeks, but sensitisation late in pregnancy is unlikely to result in HDN requiring treatment.

Quantification of the anti-D has simplified interpretation of positive antibody screens. Severe fetal anaemia is not expected at <4 IU/ml ( ) and is rare (0.25%) at <10 IU/ml. Further evaluation therefore is warranted at levels >4 IU/ml. Above this threshold, antibody levels have a limited role as they correlate poorly with the degree of fetal anaemia ( ), although a rising level suggests an increase in severity. Prior to non-invasive methods of assessment, a level >15 IU/ml indicated the need for invasive assessment of fetal anaemia. No such similar cut-off levels apply to anti-c, Kell and Fy ^a , and close monitoring is required in these cases, even with low titres.

When the fetus has been shown to be Rh-positive, the maternal antibody concentration should be checked every 2–4 weeks. Monitoring is now primarily with assessment of blood velocity in the MCA (see below). Additional ultrasonographic assessments include measurement of placental thickness, umbilical vein diameter, spleen and liver size ( ; ; ) or Doppler assessment of velocities in the descending aorta and ductus venosus ( ; ; ), although none have been shown to be as reliable as MCA Doppler measurements in predicting the degree of anaemia. While the demonstration of fetal ascites indicates severe anaemia (PCV <15%, Hb <4 g/dl) in the mid-trimester, ascites only actually develops in two-thirds of fetuses with an Hb <4 g/dl ( ). Anaemia of this degree is not associated with hypoxaemia, but is associated with increased fetal lactate levels, suggesting tissue hypoxia ( ). This is unlikely to cause developmental delay, but should be corrected as soon as possible by transfusion.

Determination of paternal zygosity and fetal genotype

Following the identification of antibodies in the maternal circulation, determination of the paternal Rh status is important. Approximately 15% of the UK population are RhD-negative, and, of the positive fathers, 56% are heterozygous for the D-gene, with a 50% chance of passing this on to the fetus. In cases of proven paternal heterozygosity, fetal antigen status should then be determined. Recent advances, since the identification of the Rh gene in 1991 ( ), have allowed fetal antigen status to be determined, firstly from amniotic fluid ( ) and now from maternal blood by identification and analysis of free fetal DNA ( ). This latter technique is known as non-invasive prenatal diagnosis (NIPD) and has replaced invasive testing in routine clinical service in the UK ( ; ). At present NIPD is only offered to women with RhD antibodies who are at risk of pregnancies complicated by HDN. It has the potential use for routine screening in all RhD-negative women, and studies are currently underway to evaluate its performance as a screening test ( ; ). Concerns of persistence of free fetal DNA from one pregnancy to the next have been largely discounted by the demonstration of its rapid clearance from the maternal circulation soon after birth.

Non-invasive assessment for fetal anaemia

Recent advances in Doppler imaging have resulted in a shift in practice away from early invasive assessment (and hence an avoidance of increased maternal sensitisation). Attention has focused on the Doppler velocimetric assessment of the MCA to predict fetal anaemia ( Fig. 9.12 ). The rationale is that the MCA responds quickly to hypoxaemia owing to the strong dependence of the brain on oxygen and also reflects the hyperdynamic circulation associated with anaemia. Several groups have shown a strong negative correlation between either peak systolic velocity or mean velocity and fetal Hb or haematocrit ( ; ; ). In the largest series, the peak systolic velocity in the MCA predicted moderate or severe anaemia without hydrops in 100% of cases, with a false-positive rate of 12% ( ). The demonstration of increased velocity in the MCA indicates a strong likelihood of fetal anaemia. At this point, invasive testing is indicated. Use of MCA monitoring compared with more traditional methods of monitoring appears to allow the first invasive procedure to be performed later in gestation without compromising fetal wellbeing ( ).

Colour Doppler appearances of the circle of Willis and the middle cerebral arteries.

Fetal blood transfusion

FBS allows direct assessment of fetal PCV and Hb, and permits transfusion to be performed at the same procedure if anaemia is detected. However, as FBS has a loss rate, and provokes fetomaternal haemorrhage and thus increases antibody levels in 70% of procedures in which the placenta is transgressed ( ), it is avoided until MCA Doppler measurements predict significant anaemia.

IVTs are now given by ultrasound-guided FBS ( ; ; ). The decision to administer an IVT is based on the PCV or Hb at FBS. Most use an Hb of less than 2 sd for the gestation ( ) as an indication for transfusion; some use an absolute haematocrit below 30%. The needle tip is kept within the umbilical vein and fresh Rh-negative packed cells compatible with the mother are infused at 10–15 ml/min. The fetal heart rate and flow of infused blood are monitored on ultrasound to guard against inadvertent needle dislodgement and cord tamponade. The volume transfused is determined by consideration of the estimated fetoplacental volume and the fetal and donor PCV ( ) or Hb ( ), according to published nomograms. The PCV is rechecked after transfusion, and, if less than the desired 40–45%, a further increment is given.

The timing of the second transfusion should be based on the pre- and posttransfusion fetal Hb from the first transfusion, and from serial MCA Doppler velocities. After the second transfusion, the rate of fall in PCV in the fetus may be determined (when the PCV is estimated to have fallen to 20–25%) and the timing of subsequent transfusions arranged, although these again are often modified by the MCA Doppler velocities. Kleihauer testing of fetal samples indicates that erythropoiesis is usually completely suppressed after two or three transfusions ( ). As the donor blood in the fetal circulation is not susceptible to immune destruction, the rate of fall in PCV declines with increasing transfusions and thus the interval between procedures can be increased. The same principles are used in scheduling delivery between 36 and 38 weeks. Once an intrauterine transfusion has been performed, the timing of delivery will depend on when the last transfusion was performed and how many days it is likely to take for the Hb to fall to a level about 2 sd below the mean. Intrauterine transfusions are not usually performed after 36 weeks. In a woman with no previous history but an antibody level above 4 IU/ml (or 1 : 16 titre), the delivery can be planned at 37–38 weeks of gestation.

Survival

With serial IVTs, survival rates of 78–95% have been achieved in severely affected fetuses ( ; ). One series has reported that, in almost 600 IVTs, intact survival was 98% in those commenced in non-hydropic fetuses at greater than 24 weeks, and 70% in hydropic fetuses less than 24 weeks’ gestation ( ). There are few follow-up studies focusing on morbidity, but one published study suggested that, when prematurity is removed, long-term morbidity secondary to intrauterine transfusion is low ( ). Therefore, the prognosis is now extremely optimistic. Fetal mortality correlates inversely with gestational age, and operator experience is undoubtedly also of importance.

Alloimmune thrombocytopenia ( Ch. 30 )

Perinatal alloimmune thrombocytopenia complicates at least 1 : 3500 births, with intracranial haemorrhage (ICH) affecting 10–20% ( ). Maternal antiplatelet antibodies cross the placenta, in a situation analogous to Rh disease. The consequent fetal thrombocytopenia may be profound, with a risk of spontaneous ICH in utero, particularly in the third trimester ( ; ; ), but it may occur as early as 18 weeks’ gestation. The human platelet-specific antigens are biallelic polymorphisms which involve the platelet surface glycoproteins. Recently, a nomenclature of human platelet antigens (HPAs) has been developed ( ). At present, HPAs 1–5 have been recognised. The genetic basis for these five HPAs has been identified, all involving a single point mutation in the DNA coding for the glycoproteins involved ( ). The most common is HPA 1, which has a high-frequency (85%) or a low-frequency (15%) antigen. Only 2% of women will be homozygous for the low-frequency ‘b’ allele, and thus at risk of developing antibodies and alloimmunisation. Fortunately, the actual occurrence of alloimmunisation is much rarer than this (0.06%) ( ) because the development of antibodies is dependent on the human leukocyte antigen (HLA) type. HLA-Drw52a and HLA-Dr3 are most commonly associated with the development of HPA 1a antibodies. The incidence of women who were HPA 1a-negative was 2% in a screening and intervention programme of 100 000 unselected pregnant women in Norway. Anti-HPA 1A antibodies were detected in approximately 10% of women who were HPA 1a-negative. In the 161 HPA 1a-positive infants of negative mothers, there were 55 cases of severe thrombocytopenia and two cases of ICH ( ).

Autoimmune thrombocytopenic purpura

Transplacental passage of antibodies in maternal immune thrombocytopenic purpura (ITP) also produces fetal thrombocytopenia. The older literature suggested that the 50% of infants with thrombocytopenia had a risk of ICH during vaginal delivery ( ), and accordingly FBS for fetal platelet count determination prior to labour used to be performed in pregnancies complicated by ITP to decide on the mode of delivery ( ; ). More recent studies show a low incidence of severe fetal thrombocytopenia (5–20%) or infant morbidity with maternal ITP ( ; ; ), with no documented cases of antenatal ICH or ICH attributable to mode of delivery, even in cases of severe thrombocytopenia. There is no correlation between maternal and fetal platelet counts ( ); the level of platelet-associated antibody does not correlate with fetal thrombocytopenia ( ; ).

Rubella

Prenatal diagnosis of rubella infection is usually indicated following maternal exposure in the early second trimester, or where doubt exists as to whether exposure in the first trimester resulted from primary infection or reinfection. Rubella-specific IgM is detected in fetal serum by radioimmunoassay ( ; ), provided that FBS can be delayed until 21–22 weeks, when the fetal humoral response to infection becomes detectable. Even at 23 weeks occasional false-negative IgM levels have been reported ( ). To improve the accuracy of prenatal testing at this late gestation, fetal blood or other tissues are also tested by hybridisation with a cDNA probe to rubella virus ( ). Earlier in pregnancy the same technique can be used on CVS specimens ( ), although concern remains that placental infection may not indicate fetal infection ( ).

Congenital infection leads to the classic rubella triad of cataracts, congenital heart defects (most commonly pulmonary stenosis) and deafness ( ). With advancing gestation, transplacental passage reduces, with congenital infection rates of 50% in the first month falling to just 10% by 3 months ( ). In addition, the severity of the syndrome reduces with advancing gestation.

Cytomegalovirus

The most severe congenital infections are due to primary maternal cytomegalovirus infection. Primary infection is associated with a 40% transmission rate, although only 10% of those fetuses will develop long-term sequelae, mostly hearing and learning defects ( ). Among the 5–10% who are symptomatic at birth, there will be a neonatal mortality of 30%, with long-term handicap in all survivors. Infection in earlier gestation results in higher transmission and poorer outcome.

Prenatal diagnosis is based on ultrasound findings and the detection of viral particles in the amniotic fluid ( ), or, less commonly, in fetal blood. Specific IgM in fetal serum may also be diagnostic ( ). Ultrasound findings include FGR, ascites or hydrops, intracranial calcification, ventriculomegaly and bowel echogenicity. Nevertheless, 90% of infants with congenital cytomegalovirus infection remain neurologically and developmentally normal. The risk of adverse outcome is increased if there is microcephaly, intracranial calcification, ventriculomegaly or evidence of a cerebral migration disorder ( Ch. 40.8 ), and in these situations termination of pregnancy should be discussed.

Toxoplasmosis

Although, as pregnancy advances, maternal exposure leads to an increased risk of fetal infection, severity is greatest with exposure in the first trimester, with little chance of severe congenital disease after 20 weeks’ gestation. Termination of an infected fetus remains an option, although fetoplacental infection is largely treatable ( ). The aim of fetal testing is to allow optimal transplacental therapy, initially with maternal spiramycin (3 g/day) to prevent transplacental transmission, with the addition of pyrimethamine and sulfadiazine if fetal testing proves positive ( ). These two drugs are directly antiparasitic and have been shown to limit fetal damage. They are not used in the first instance when information about maternal infection is known, but rather only when fetal infection is proven, because of the potential hazards to mother and fetus. The diagnosis of fetal infection is now made with PCR analysis of amniotic fluid, which has replaced the more cumbersome multiple testing that was previously required ( ). If fetal infection is proven, the prognosis is still likely to be good, although vision may be affected later. However, intracranial signs of calcification and/or ventriculomegaly are poor prognostic signs (Berrebi et al. 1994), and termination would be offered.

Parvovirus

Human parvovirus (B19) infection is associated with increased risks of miscarriage, hydrops and IUD ( ), with an overall fetal loss rate of 9% in infected pregnancies ( ). Parvovirus is best identified in fetal blood or other tissues by dot-blot hybridisation, electron microscopy or, most commonly, PCR of B19-specific DNA ( ). The mainstay of diagnosis, however, remains maternal serology in women with appropriate clinical symptoms. Anti-B19 IgM appears in the serum at the onset of illness and remains detectable for up to 3 months. IgG response begins after 7 days and persists, probably to confer lifelong immunity. FBS is not routinely indicated in pregnancies with maternal infection because at least 90% will result in livebirths and parvovirus does not seem to be teratogenic.

It is the profound fetal anaemia secondary to an infective erythroid aplasia that accounts for the significant mortality rate. Following documented maternal infection, close fetal monitoring for signs of anaemia is indicated. Assessment of the MCA velocity for early detection of anaemia may be useful (as with monitoring of fetuses at risk of alloimmunisation) for 12 weeks postexposure, as fetal anaemia may occur 1–11 weeks after maternal infection ( ). FBS and IVT are indicated in the presence of suspected anaemia or hydrops. The incidence of coexisting fetal thrombocytopenia requiring platelet transfusion has been reported as high as 25% ( ; ). Hydrops without fetal anaemia has been documented and viral particles have been identified in the fetal myocardium, suggesting that a myocarditis may be contributory. It appears that profoundly anaemic fetuses which are salvaged by transfusion have a good prognosis, although more than one IVT may be necessary on occasion ( ; ).

Oligohydramnios

Causes of oligohydramnios in the mid-trimester include urinary tract malformations, preterm premature rupture of membranes (PPROM) and FGR, and at these gestations survival is less than 25% whatever the cause ( ; ). Conditions with a lethal prognosis, such as renal agenesis and aneuploidies, should be ruled out. Absence of the acoustic window makes inspection of fetal anatomy difficult. Transvaginal ultrasound facilitates visualisation of the renal fossae, as does colour Doppler of the renal arteries ( ). If still equivocal, invasive procedures may be required, including amnioinfusion and/or instillation of fluid into the fetal peritoneal cavity ( ). Amnioinfusion of a warmed physiological solution not only restores the acoustic window, but also allows confirmation of PPROM, especially when a dye is added ( ).

As 5–10% of fetuses in pregnancies with severe oligohydramnios will be chromosomally abnormal ( ), rapid karyotyping is carried out at the time of amnioinfusion. In the rare case of a euploid fetus with oligohydramnios, intact membranes and an intact renal tract, there may be a role for serial amnioinfusions in the prevention of lethal pulmonary hypoplasia ( ), which otherwise complicates at least 60% of cases of severe mid-trimester oligohydramnios ( ).

Polyhydramnios

The more severe the polyhydramnios, the more likely that an underlying cause will be found. Using the maximum vertical pocket, mild and severe polyhydramnios have been arbitrarily defined as a deepest pool greater than 8 cm and 15 cm, respectively ( ). The amniotic fluid index (AFI) definitions for mild and severe polyhydramnios are values outside the 97.5th centile for gestation and an AFI >40 cm, respectively ( ).

Exclusion of maternal diabetes is essential; thereafter, a detailed fetal assessment is mandatory. In one series, 11% of neonates had a structural anomaly (most notably tracheo-oesophageal fistula, oesophageal atresia, duodenal atresia or conditions that impair fetal swallowing such as arthrogryposis), which increased with increasing severity of the polyhydramnios ( ). If sonographic evaluation was normal, the risk of a major anomaly was just 1% with mild polyhydramnios, but 11% if severe. Aneuploidy was present in 10% of fetuses with sonographic anomalies and in 1% without. Overall the fetal loss rate was 4%, of which 60% had anomalies.

The increased risks of preterm labour and PPROM in polyhydramnios seem mainly confined to those with severe polyhydramnios, i.e. an AFI >40 cm or a deepest pool >15 cm ( ; ). Amniotic reduction (AR) has been used with anecdotal success in severe polyhydramnios in order to prolong gestation and relieve maternal discomfort ( ; ). Removal of relatively small volumes of amniotic fluid can restore amniotic pressure to normal ( ), but usually larger volumes are removed to limit the number of procedures that may be required. Nevertheless, removal of volumes greater than 6 litres at one time does carry the risk of precipitating abruption and/or preterm labour ( ).

Prostaglandin synthetase inhibitors have been used to reduce amniotic fluid in the past but are no longer recommended owing to potential fetal side-effects.

Twin–twin transfusion syndrome

TTTS complicates approximately 15% of monochorionic diamniotic twin pregnancies ( ), and accounts for 15–17% of overall perinatal mortality in twins ( ; ). It has classically been attributed to transfusion of blood via placental vascular anastomoses between the two fetal circulations. Vascular anastomoses are present in 96% of monochorionic placentas, and interfetal transfusion is a normal event in the majority of these pregnancies ( ). However, placentas from pregnancies affected by TTTS are characterised by two findings: first, an imbalance in interfetal transfusion is set up by the presence of unidirectional arteriovenous anastomoses, and, second, there is an absence of compensatory bidirectional arterioarterial anastomoses ( ; ).

Presentation in the mid-trimester has until recently been associated with an 80–100% perinatal loss rate, either from preterm delivery secondary to polyhydramnios or from IUD following severe growth restriction or circulatory overload ( ; ). The majority of cases complicated by TTTS have estimated fetal weight discordance in utero, and differences of <15% are unusual ( ). There is some evidence that the prognosis worsens with an increasing difference in abdominal circumference measurements ( ), cumulatively reflecting the severity of overload/early hydrops in the recipient and growth restriction in the donor.

TTTS presents in the second trimester with discordant amniotic fluid volume ( Fig. 9.13 ) ( ). The donor twin becomes oliguric, with consequent oligohydramnios, hence appearing enshrouded within its membrane and finally stuck to the uterine wall. In addition, neither the urinary bladder nor the stomach is usually visible, once anhydramnios is present. The growth velocity of the donor may follow the asymmetrical pattern seen with fetuses affected by FGR. Furthermore, it may also exhibit abnormal umbilical and MCA velocities, with absent/reversed flow in the former and redistribution in the latter ( ).

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. Basic genetics
2. Pharmacology
3. Ophthalmology
4. Coagulation values in term and preterm infants

Stay updated, free articles. Join our Telegram channel

Join